My current research agenda falls into two main areas. The first area is my focus on a population-based approach to examining the quality of care for women with breast cancer. The second area is the evaluation of online communication and service tools to improve the quality of cancer early detection initiatives in primary care settings. My research agenda over the next five years will extend my current work in several ways. First, I plan to expand my current SEER-based population research program to address the quality of care for women with breast cancer. This will include: 1) additional population-based studies of determinants and outcomes of other potentially marginalized patient populations;and 2) population-based intervention studies to improve communication, treatment processes and outcomes for women with breast cancer. Second, I will build and evaluate novel online communication and service tools for patients and providers to improve cancer screening processes and outcomes in primary care settings. The goals of my mentoring plan are 1) to increase the quantity and quality of NIH investigator-initiated proposals in oncology health services research from the University of Michigan;2) to foster interest in oncology-related health services research by fellows and junior investigators through my leadership on clinical training programs at the University of Michigan;and 3) recruit new investigators to the medical school focused on health services research in oncology. In order to address the goals above I will lead the following initiatives: 1) to obtain intramural and extramural funding to recruit and support clinician investigators and non-clinician research scientists in oncology health services research;2) to personally mentor these investigators towards a successful career;3) to develop and direct educational programs to support this area of research;and 4) to coordinate efforts in this area across professional schools at Michigan. The structure, process and outcomes of my personal mentoring program are described in detail in the application.